Search

Background: Oscillatory activity in the beta range, in human primary motor cortex (M1), shows interesting dynamics that are tied to behaviour and change systematically in disease. To investigate the pathophysiology underlying these changes, we must first understand how changes in beta activity are caused in healthy subjects. We used the Dynamic Causal Modelling framework to develop a microcircuit model to explain mesoscopic dynamics in motor cortex. This model was applied to explain differences in measured oscillatory characteristics between healthy controls and stroke patients. The model was tested for robustness of use in EEG, and applied to pharmacological datasets to investigate differences in effective connectivity at the mesoscopic level under different pharmacological conditions. Aims: To investigate the laminar interaction underpinning beta-oscillatory dynamics in humans, in vivo, non-invasively. To this apply methodology to stroke patients to elucidate any differences at the mesoscopic level between these two key groups. To investigate whether such a technique was clinically feasible with more readily available research tools such as EEG. To investigate mesoscopic motor cortical dynamics under pharmacological influence, in humans, in vivo. Data acquisition and analysis: Most data acquisition was performed using a 275 CTF MEG scanner, with some data acquisition being performed using a standard 32 lead EEG headset. These data were taken and subject to rigorous analysis, that utilised dynamic causal modelling, Bayesian model comparison, as well as several signal processing and head modelling procedures that is outlined in detail in chapter 2. All analysis was done using the SPM 12 suite designed for neuroimaging analysis in MATLAB. Conclusions: We were able to develop a canonical microcircuit model for M1, and show it had significantly more model evidence than previous CMC models in explaining data from motor cortex. We applied this model to healthy controls to show the laminar interactions underpinning beta-oscillations in humans, in vivo. We were then able to apply this model and characterise laminar differences between healthy controls and stroke patients, as well as propose a novel mechanism for the origins of movement related beta desynchronization. We were able to show that our technique remained robust when applied to more clinically appropriate EEG. We were able to show significant differences in effective connectivity in different pharmacological states, that corresponded to differences observed in measured oscillatory data between groups who were given tiagabine and controls.